1. Field of the Invention
The present invention relates to an image sensor, and more particularly, to an image sensor with a shield electrode for solving carrier cross-talk problems.
2. Description of the Prior Art
The image sensor such as complementary metal oxide semiconductors (CMOS) or charge coupled device (CCD) is a silicon semiconductor device designed to capture photons (light) and convert them into electrons. Electrons, once converted, must then be transferred and converted again to voltage which can be measured and turned into digital data. The hydrogenated amorphous silicon (α-Si:H) based image sensor stacking on CCD or CMOS elements has been studied to pursue the advantages over the conventional CCD or CMOS image sensor as below. The high fill factor brought by its stacking structure will provide the full of pixel area to be available for photo sensing, thereby achieving the high quantum efficiency in conjunction with the direct energy transition of α-Si:H material. However, this type of sensor has been suffered by cross-talk, image lag, and dark leakage signal problems in the past study. In particular, the problem of carrier cross-talk across adjacent pixels causes the serious resolution and uniformity degradation at the photo response, also brings the color cross-talk over the pixels that produces the poor color fidelity. Besides, the image lag problem is mostly caused by low carrier velocity due to its deep traps or field emission based carrier transportation mechanism through the α-Si:H material, in which the image tailing appears from the bright spot over the frames in the motion picture. In particular, it is impossible to reproduce the color fidelity at low signal level when the sample having an image lag problem, since the full amount of signal from a pixel can not be read out in just one shot frame. Finally, the dark leakage problem is mainly caused by the injection of hole or electron from metal electrodes to the p-type layer (p-layer) or n-type layer (n-layer) of the photo conductive layer by tunneling, which generates the cosmic noise in the dark scene. Thus, these three major problems should be improved to compete with the conventional silicon based CCD or CMOS image sensors for the better picture quality.
The current technology utilizing α-Si:H has been developed to comprise with the following materials: 1) the transparent metal such as ITO (Indium Tin Oxide); 2) the heavily boron doped p-layer composed of hydrogenated amorphous silicon carbide (α-SiC:H) to collect photo generated holes formed on an intrinsic layer (i-layer) to contact with the ITO; 3) the intrinsic α-Si:H layer (i-layer) which is mostly contributed as the photonic electron-hole pairs generation layer, 4) the heavily phosphorus doped n-layer composed of hydrogenated carbon doped amorphous silicon working as the receptor for electrons from i-layer to the metal pixel electrodes; and 5) the metal pixel electrode, under the n-layer connecting to the transistor, and vertically stacked above the CMOS circuitry implemented on a silicon substrate. FIG. 1 shows the band diagram of the p-i-n hetero-junction structure having an i-layer/n-layer interface as described above. The charge to voltage conversion gain is mainly determined by the sensing capacitance in which it can be minimized by the thickening i-layer.
For achieving the higher quantum efficiency in the α-Si:H based i-layer, the photo-conductivity and light absorption should be improved with longer minority carrier lifetime and higher carrier mobility for large depth by optimizing hydrogen content. Meanwhile, the heavily boron doped p-layer under the ITO layer can be changed to a CH4 based α-SiC:H layer forming the hetero junction with the α-Si:H based i-layer. It is effective to enhance the transparency due to the larger optical band gap (Eopt) of SiC, also to prevent the electron emission by tunneling effect from the ITO layer to the p-layer by widening the energy band gap for the dark leakage suppression. Furthermore, the α-SiC:H material is also applicable for the n-layer to avoid the lateral carrier cross-talk across the pixels by reducing the conductivity of the n-layer between the pixel electrodes. Also it is effective to block the hole emission from the pixel electrodes formed with titanium nitride (TiN) to the n-layer as the same case as the electron tunneling to the p-layer. However, the α-SiC material brings the dark signal and image lag problems caused by its high density of deep traps.
Furthermore, there occurs a more serious problem at the pixel electrode corners have a boundary with the n-layer since the electric field strength is locally concentrated to bend the energy band at the pointed edge as shown in FIG. 2, wherein the transition probability of holes by tunneling becomes higher to increase the dark leakage current at the reversed biased condition. Besides, the n-layer composed of α-SiC is also effective for adhesion on pixel electrode formed with TiN without peeling-off, which is also occurred at the pointed electrode corner by the tensile force. However, the stressed α-SiC film on the corners of the pixel electrode may have high density of traps to cause the pixel defects as well as the image lag worse, as shown in FIG. 2.
Referring to FIGS. 3(a)-4, FIG. 3(a) is a cross-sectional schematic diagram of an image sensor utilizing stacked p-i-n layer structure according to the prior art, FIG. 3(b) is a proposed equivalent circuitry of the image sensor of FIG. 3(a), and FIG. 4 is a schematic band diagram of the pixel electrode and pixel electrode gap shown in FIG. 3(a). The prior-art image sensor 10 comprises a plurality of pixel circuits (not shown) and an isolation film 24 on a substrate (not shown), a plurality of pixel electrodes 12 on the pixel circuits and the isolation film 24, a photo conductive layer 14 on the pixel electrodes 12, and a transparent electrode 16 on the photo conductive layer 14, wherein the photo conductive layer 14 comprises an n-layer 18, an i-layer 20, and a p-layer 22 from bottom to top, which constitute so-called stacked p-i-n layer structure.
The photoconductive layer 14 is introduced in conjunction with four capacitive components such as Cpd, Csub, C1 and C2. These capacitive components are oriented from the node point at the n/i-layer interface in the middle of the pixel electrode gap as shown in FIG. 3(a). Here, Cpd is represented as the capacitance component for the transparent electrode 16, formed with indium tin oxide (ITO), Csub is represented as the capacitance component for the p-type silicon substrate (not shown) through silicon oxide (SiO2) based isolation films 24, while C1 and C2 are coupled with the adjacent metal pixel electrodes 12, respectively. This device structure of the prior-art image sensor 10 can be assumed as the flipped n-channel metal insulator semiconductor field effect transistor (MISFET) 30 where both the source and the drain are connected to two pixel electrodes 12 individually, as shown in FIG. 3(b), while the substrate bias is supplied from the transparent electrode 16 to the p-layer 22, and then the silicon substrate at the grounded level will be regarded as the gate of the assumed MISFET device 30 having the gate capacitance of Csub.
Since the Cpd and Csub would be not large enough to compare with C1 or C2 due to their aspect ratios in the actual device structure, the potential distribution between the adjacent pixel electrodes 12 is strongly modulated with the bias voltage, which is near power supply voltage level applied on the adjacent pixel electrodes, by the lateral 2-dimentional effect. Therefore, the channel potential profile of the flipped MOSFET 30 is highly pulled up by the coupling through C1 and C2 from the pixel electrodes 12, while Cpd or Csub cannot sustain the potential in the pixel electrode gap region to lower the level. Thus, the electro-static potential barrier height for electrons becomes lower than that of 1-dimentional approximation, thereby generating the cross-talk current across the pixels as shown in FIG. 4.
FIG. 5 shows the energy band diagram of the device structure for both the pixel electrode and the electrode gap region shown in FIG. 3(a) vertically, wherein the electrode gap means the spacing between the adjacent pixel electrodes 12, and the electron channel layer is located at the i-layer 20/n-layer 18 interface. Since materials of i-layer 20 and n-layer 18 have the different energy conduction band levels which form the hetero-junction band, the electrons are mostly accumulated and stored at this interface to form the electron channel layer, and photo-generated electrons will be flowed over the n-layer 18 conduction band to the pixel electrodes 12. On the other hand, the horizontal carrier path across the adjacent pixel electrodes 12 will be formed through the i-layer 20/n-layer 18 interface to cause cross-talk effects as previously explained in FIG. 3(a) and FIG. 4. Furthermore, the potential barrier in the pixel electrode gap region is lowered from the 1-dimentioanl barrier height as shown in FIGS. 4 and 5. Considering from the proposed MISFET model as shown in FIG. 3(b), this cross-talk may be suppressed by thinning i-layer 20 or applying higher voltage on the pixel electrodes 12 to utilize the enhancement of body effect. Although it is obvious that thicker i-layer has rich quantum efficiency, the image sensor with a thicker i-layer has lower immunity for cross-talk due to its weak body effect demonstrated by the potential distribution between the pixel electrode gap region. The sensitivity and the color balance of the prior art image sensor are sacrificed by using a thin i-layer 20, and a higher pixel electrode voltage cannot satisfy the electrical specification for power supply.